Crystengcomm Accepted Manuscript

CrystEngComm Accepted Manuscript

This is an Accepted Manuscript, which has been through the Royal Society of Chemistry peer review process and has been accepted for publication.

Accepted Manuscripts are published online shortly after acceptance, before technical editing, formatting and proof reading. Using this free service, authors can make their results available to the community, in citable form, before we publish the edited article. We will replace this Accepted Manuscript with the edited and formatted Advance Article as soon as it is available.

You can find more information about Accepted Manuscripts in the Information for Authors.

Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal’s standard Terms & Conditions and the Ethical guidelines still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains.

www.rsc.org/crystengcomm Page 1 of 12 PleaseCrystEngComm do not adjust margins

CrystEngComm

ARTICLE

Isoniazid cocrystallisation with dicarboxylic acids: vapochemical, mechanochemical and thermal methods

Received 00th January 20xx, a† a a Accepted 00th January 20xx I. Sarceviča, A. Kons and L. Orola

DOI: 10.1039/x0xx00000x Cocrystallisation with a series of related compounds allows for the exploration of trends and limitations set by structural differences between these compounds. In this work, we investigate how the length of a dicarboxylic acid influences the www.rsc.org/ outcome of cocrystallisation with isoniazid. We have performed a systematic study on the mechanochemical, thermal and

solvent vapour-assisted cocrystallisation of aliphatic dicarboxylic acids (C 3–C10 ) with isoniazid. Our results demonstrate that the rate of mechanochemical and vapour-assisted cocrystallisation depends on acid chain length and shows alternation between odd- and even- chain acids. The results of thermal cocrystallisation showed that eutectic melting temperatures of isoniazid–dicarboxylic acid mixtures follow the same trend as do the melting points of dicarboxylic acids.

13,22 formation of a cocrystal . The thermal method is a fast Manuscript Introduction method compared to crystallisation from solution. Alongside mechanochemistry, thermal cocrystallisation also usually gives As a way to improve the characteristics of materials, cocrystal 11,14 1–4 good results in cocrystal screening and can be efficiently form has recently became of a large interest . The possibility used to identify systems that can form cocrystals. to tune the properties of compounds via cocrystallisation has 23 2,3,5 Spontaneous cocrystallisation is also possible and is usually been shown and has been used to develop new materials 24,25 6,7 8 facilitated by moisture . Although the promotional effect of for non-linear optics , semiconductors and pharmaceutical 24,25 26,27 1–3 the vapour of water or organic solvent on applications . Various methods based on crystallisation from solution 9–12 or from melt (thermal cocrystallisation) 11,13,14 , cocrystallisation is known and has been investigated by several 15–18 groups, little practical advantage of this knowledge has been

mechanical treatment etc. are used for cocrystal Accepted preparation. From these, the mechanochemical gained until now. Perhaps the lack of appreciation for solvent cocrystallisation is typically recognised to give the best vapour-assisted cocrystallisation as a cocrystal screening results 11 , moreover, liquid-assisted grinding is more efficient method stems from the difficulty in predicting whether it will compared to neat grinding (no solvent used). In screening be useful in the system of interest. Consequently, this method experiments that employ and compare different mostly has been neglected in systematic screenings and only cocrystallisation methods, solvent–drop grinding has been individual cases of vapour-assisted cocrystallisation have been shown to produce the largest number of cocrystals 11 . In reported. Our results, however, imply that vapochemistry addition, mechanochemistry allows inter-conversion between could be beneficial for cocrystallisation of small molecules. In cocrystals 19 of different stoichiometry and polymorphism order to evaluate the usability of the vapochemical approach control 20,21 . Further benefits of the mechanochemical method in each system, it is important to understand the limitations include minimal use of solvent, high yields and easiness to set by the molecular parameters of cocrystal ingredients and perform 17 . by solvent properties. One of the possible limitations could be The thermal cocrystallisation method is based on the size of molecules, as the movement of large molecules understanding that cocrystals can form from the eutectic melt. would be hampered. The presence of thermal effects (in the thermal analysis data) There are several explanations available for the promotional effect of water vapour on cocrystal formation. In some cases,

corresponding to eutectic melting and crystallisation implies CrystEngComm the adsorbed water has been shown to partially dissolve cocrystal ingredients, allowing cocrystallisation from a saturated solution 24,25 . In other examples water serves as a plasticizer and therefore promotes cocrystallisation 28 . Various solvents are expected to influence cocrystallisation differently in accordance with their properties. Therefore, solvent vapour- assisted cocrystallisation is also expected to offer polymorph control, similar to crystallization from different solvents 29 .

This journal is © The Royal Society of Chemistry 20xx J. Name ., 2013, 00 , 1-3 | 1

Please do not adjust margins PleaseCrystEngComm do not adjust margins Page 2 of 12

ARTICLE Journal Name

A commonly used set of cocrystal formers is the series of stoichiometry, although malonic acid is an odd-chain aliphatic dicarboxylic acids 30–33 . The choice of these dicarboxylic acid. In our mechanochemical, thermochemical dicarboxylic acids as cocrystal formers is advantageous in and vapochemical cocrystallisation experiments, we used many aspects. They easily form cocrystals with hydrogen these known stoichiometries. acceptor compounds; their crystal structures 34 and physicochemical properties 35–39 are described in the literature. Mechanochemical cocrystallisation In addition, dicarboxylic acids offer a set of related compounds To investigate the effect of molecule size on the rate of with similar molecular structures, and therefore render the mechanochemical cocrystal formation, milling experiments interpretation of results easier. Most physicochemical were performed for isoniazid and dicarboxylic acid (C 3−C10 ) properties of dicarboxylic acids follow an interesting trend of mixtures. The conversion (relative amount of cocrystal in a alternation between odd- and even-chain acids. For example, sample) was acquired by Rietveld analysis of the diffraction the solubility 35,36 and saturated vapour pressure 37,38,40 of odd- patterns of milling products. Inspection of PXRD patterns of chain acids are higher than for even-chain acids. The melting milling products showed that they contain cocrystal phases points 34 and vaporisation enthalpies 40 of odd-chain acids are with the known crystal structures 41 of isoniazid–dicarboxylic lower, compared with those of even-chain acids. These trends acid cocrystals, and these structures were used in Rietveld are commonly explained by the twisted conformations of odd- refinement. It should be noted that some peak broadening and chain acids in their crystal structures 34,37 , leading to lower minor background changes were observed in the powder X-ray stability of these structures compared with crystal structures diffraction (PXRD) patterns of milling products, implying a of even-chain acids. Interestingly, the cocrystals of dicarboxylic possible presence of amorphous phase. Results of the acids also often show the odd-even alternation effect in their quantitative analysis should therefore not be seen as absolute physicochemical properties 30–33 . values, but rather as an indication of the extent of Here we report a series of experiments that demonstrate the cocrystallisation (the amount of the cocrystal relative to the Manuscript vapour-assisted, thermal and mechanochemical amounts of crystalline isoniazid and acid). Examples of PXRD cocrystallisation of an anti-tubercular drug isoniazid with patterns of milling products are available in Figures S1–S8 in aliphatic dicarboxylic acids (C 3–C10 ) as cocrystal formers ESI. (Scheme 1). Isoniazid was chosen as a model compound for Figure 1 presents cocrystallisation results for isoniazid and our experiments as it is known to form cocrystals with all C 3–C8 dicarboxylic acid milling experiments performed for 5 and 15 dicarboxylic acids 41,42 and many other acidic compounds 41,43–53 . minutes under the same conditions. A comparison of the Organometallic complexes 54–58 and salts 59,60 with inorganic results shows that longer mechanical treatment returned a anions of isoniazid are also known. Our experiments reveal higher amount of cocrystal. In the case of isoniazid–glutaric how the length of the acid and its physicochemical properties acid sample, however, the conversion difference was small due

determine trends in cocrystal formation with isoniazid. to partial amorphisation and formation of an unidentified Accepted phase. Cocrystallisation of isoniazid and malonic acid proceeded rapidly compared with other dicarboxylic acids, and almost complete conversion was achieved after 5 minutes of milling. We observed that isoniazid–malonic acid cocrystal starts to form already during sample preparation under laboratory conditions, and milling accelerates the reaction.

Cocrystallisation with C 4–C7 dicarboxylic acids proceeded considerably more slowly than cocrystallisation with malonic acid and the conversion after 5 minutes did not exceed 40% in any of these samples.

CrystEngComm

Scheme 1 . Structural formulas of isoniazid and C 3-C10 dicarboxylic acids.

Results and discussion It has been found previously 41,42 that isoniazid forms 1:1 cocrystals with odd-chain dicarboxylic acids and 2:1 cocrystals with even-chain dicarboxylic acids. An exception is the isoniazid–malonic acid cocrystal, which has the 2:1

2 | J. Name ., 2012, 00 , 1-3 This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins Page 3 of 12 PleaseCrystEngComm do not adjust margins

Journal Name ARTICLE

Fig. 1 The relative amount of cocrystal in isoniazid and dicarboxylic acid milling (20 Hz; 5 (red) and 15 (blue) minutes) products as a function of acid chain length.

In order to evaluate the time necessary to achieve complete Fig. 2 Comparison of simulated powder X-ray diffraction patterns of isoniazid, adipic acid, isoniazid–adipic acid cocrystal form I and isoniazid–adipic acid cocrystal form II. conversion, we performed experiments with extended milling time. These results showed that the conversion time for Traces of isoniazid were observed in the powder X-ray isoniazid–succinic acid cocrystal is approximately 200 minutes diffraction patterns of the product even after 260 minutes of

and the product contains pure cocrystal phase (Figure S2 in Manuscript milling, but the presence of the adipic acid was not detected. ESI). The co-milling of isoniazid and glutaric acid required The formation of isoniazid–pimelic acid cocrystal required approximately 150 minutes, and the product contained approximately an hour; the final product contained traces of residue of isoniazid but not glutaric acid (Figure S3 in ESI). unreacted isoniazid (Figure S5 in ESI). Significant amorphisation for this sample was also observed. The cocrystallisation of longer-chain acids (suberic acid, azelaic The diffraction reflections corresponding to an unidentified acid, sebacic acid) was not achieved by milling at 20 Hz for up phase observed during the initial stage of treatment later to 15 minutes. However, after an hour of co-milling isoniazid disappeared implying that this phase is possibly an with suberic acid, formation of the cocrystal took place (Figure intermediate. For the isoniazid–adipic acid sample, formation S6 in ESI). Even longer milling, up to 200 minutes, allowed of two polymorphic forms of the cocrystal (I and II) was obtaining the cocrystal with only traces of isoniazid (and again observed depending on the milling time (Figure S4 in ESI). Accepted – no suberic acid). The isoniazid–azelaic acid cocrystal in Initially, the form I (the known polymorph of isoniazid – adipic mechanochemical experiments was not obtained even after acid cocrystal) was acquired but later it converted to a new 300 minutes of milling (Figure S7 in ESI). In the mixture of polymorphic form II. This cocrystal also formed in milling isoniazid and sebacic acid, cocrystal formation was observed experiments if the milling frequency was 30 Hz and the milling after 200 minutes of milling but it proceeded very slowly and time exceeded 30 minutes. The isoniazid–adipic acid cocrystal even 300 minutes of mechanical treatment was not sufficient polymorph II could also be prepared by pre-milling the 2:1 to achieve complete conversion to cocrystal (Figure S8 in ESI). physical mixture of isoniazid and adipic acid and then keeping Since amorphisation to some degree was observed for most of this mixture in the presence of water vapour. A comparison of the samples, additional experiments were executed to the diffraction patterns simulated form crystal structures of evaluate the changes related to the crystallinity of the acids both isoniazid–adipic acid polymorphs and cocrystal during milling. The results of these experiments showed ingredients is shown in figure 2. increase in the diffraction peak width with increasing acid

chain length. The peak broadening can be related to partial amorphisation of the acid as well as the reduction of particle size.

The differences in the mechanochemical reaction rate CrystEngComm between acids can be related to several factors including the stability of the lattice of the acid, mobility of molecules and partial amorphisation of acids during milling. An alternation between formation rates of isoniazid cocrystals with odd- and even-chain acids was observed with the malonic, glutaric and pimelic acid cocrystals forming faster, compared with cocrystals with adjacent even-chain acids. Similar odd-even alternation effects have been observed for many physicochemical properties (solubility, melting points, etc.) of

This journal is © The Royal Society of Chemistry 20xx J. Name ., 2013, 00 , 1-3 | 3

Please do not adjust margins PleaseCrystEngComm do not adjust margins Page 4 of 12

ARTICLE Journal Name dicarboxylic acids 35,36 implying that the stability of even-chain The eutectic temperatures of isoniazid −dicarboxylic acid acid crystal structures is higher than that of odd-chain mixtures show similar pattern to melting points of pure acids. dicarboxylic acid crystal structures. The disintegration of the The eutectic temperatures and acid melting points are higher less stable odd-chain diacid crystal structures therefore for isoniazid mixtures with even-chain acids than for odd-chain requires less energy leading to an easier cocrystallisation. acids. In the DTA curve of the isoniazid −azelaic acid 1:1 The outcome of mechanochemical cocrystallisation also mixture, only one peak was observed implying that depends on mass transfer processes 61 and the rate of cocrystal recrystallisation from melt had not taken place. formation is therefore affected by the ability of molecules to The values of cocrystal melting points are between the values move. Small molecules, which generally tend to be more of eutectic temperatures and the melting points of acids, with mobile, cocrystallise faster than larger molecules. The effect of the exceptions of glutaric acid and pimelic acid. The melting the size of molecules is demonstrated by the isoniazid points of isoniazid–glutaric acid cocrystal and isoniazid–pimelic cocrystals with suberic and sebacic acid that required much acid cocrystal are slightly higher compared to melting points of longer time to form compared with other cocrystals with pure acids. Interestingly, the DTA curve of isoniazid–malonic shorter acids. acid sample did not show a distinct endothermic effect Interestingly, the amorphisation during mechanochemical corresponding to eutectic melting of the cocrystal. A broad milling does not correlate with the amorphisation of pure acids exothermic effect, thought to be the recrystallization of the (Figures S9-S16 in ESI). Considerable peak broadening and cocrystal occurred at 65 °C. background changes were observed for isoniazid–glutaric acid Thermal analyses do not provide information on the structure and isoniazid–succinic acid samples although these acids did of cocrystallisation products; therefore, we used variable not show high amorphisation when milled separately. The temperature powder X-ray diffraction experiments (VT-PXRD) amorphisation of these samples could therefore be related to to obtain more information on the products of thermal partial eutectic melting during the milling process. cocrystallisation. The VT-PXRD analysis implied that the Manuscript isoniazid cocrystals of malonic, succinic, glutaric, pimelic, Thermal cocrystallisation suberic and sebacic acids obtained using the thermal method Crystallisation from eutectic melt is another popular way of were the same as reported in the literature 41 and as obtained preparing cocrystals, and it is often used as a fast and efficient mechanochemically, although the samples of isoniazid– screening method 13,22 . The odd-even alternation between malonic acid and isoniazid–glutaric acid cocrystals also values of melting points of dicarboxylic acids 34 and their contained an unidentified phase (Figures S17–S23 in ESI). The cocrystals 30,33 encouraged us to study the relation between the thermal cocrystallisation products of odd-chain acids (malonic length of the dicarboxylic acid and the eutectic temperatures acid, glutaric acid, pimelic acid) also contained residues of in mixtures with isoniazid. The eutectic temperatures were isoniazid as the dicarboxylic acid had vaporised at elevated

acquired from differential thermal analysis (DTA) of physical temperature. The formation of isoniazid–azelaic acid cocrystal Accepted mixtures of isoniazid and dicarboxylic acids and are shown in via eutectic melting was not observed. Figure 3 along with the melting points of pure diacids and their The isoniazid–succinic acid cocrystal with a known crystal cocrystals with isoniazid. The melting points of cocrystals were structure 41 , further referred to as polymorph I, formed initially also taken from DTA results and corresponded to the at 124 °C but underwent a phase transition to form a new recrystallisation product of the eutectic mixture. polymorph (form II) when kept at this temperature for >3 h. PXRD patterns showing the conversion of the physical mixture of isoniazid and succinic acid to cocrystal and the following polymorphic transition are presented in Figure 4.

CrystEngComm

Fig. 3 Melting points of pure dicarboxylic acids, their cocrystals with isoniazid and the eutectic temperatures of isoniazid–dicarboxylic acid physical mixtures.

4 | J. Name ., 2012, 00 , 1-3 This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins Page 5 of 12 PleaseCrystEngComm do not adjust margins

Journal Name ARTICLE

between odd-chain and even-chain diacid crystal lattices. These differences result in alternating trends in the saturated vapour pressure 37–40 and sublimation enthalpy 37 of the acids. As the sublimation is easier and the saturated vapour pressure is higher for odd-chain acids, the cocrystallisation of these acids to isoniazid proceeds faster. Furthermore, sublimation may increase the disorder in the surface layer 64 of crystallites, thus raising their reactivity. For the longer dicarboxylic acids (suberic acid, azelaic acid and sebacic acid) cocrystallisation within the experimental time frame (24 h) was not observed. An exception was the isoniazid–suberic acid cocrystal that formed only in the presence of ethanol vapour. This exception is explained by adsorption of ethanol onto the surface of the sample, resulting in the formation of a liquid layer. It is thought that crystallisation of a cocrystal from this saturated solution had taken place. Since the experimental time frame of 24 h in the Fig. 4 Powder X-ray diffraction patterns showing isoniazid–succinic acid cocrystal presence of water, acetonitrile, ethyl acetate, chloroform and polymorph transition at 124 °C. Blue arrows indicate the disappearance of toluene vapour was not sufficient to obtain cocrystals of long diffraction peaks characteristic to isoniazid–succinic acid cocrystal form I and the red arrows indicate the appearance of diffraction peaks characteristic to acids, we extended the time to 1 month. The prolonged isoniazid–succinic acid cocrystal form II. experiment afforded the formation of isoniazid–suberic acid cocrystal in the presence of acetonitrile and ethyl acetate Manuscript The thermal cocrystallisation of isoniazid and adipic acid vapour. Neither isoniazid–azelaic acid nor isoniazid–sebacic returned the polymorph II. The thermal cocrystallisation in the acid cocrystal were obtained in vapochemical cocrystallisation mixture of isoniazid and sebacic acid was slow and only partial experiments. conversion was observed after 6 h of storage at 113 oC. The effect of the solvent vapour can be explained by its adsorption on the surface of the sample followed by Solvent vapour-assisted cocrystallisation interaction with molecules in the surface layer of crystallites. In the previous examples, cocrystal formation was achieved by Such interactions can be specific hydrogen bonds, π-π applying mechanical or thermal energy to the mixture of interactions, Van der Waals forces etc. and they depend on the cocrystal ingredients; however, spontaneous cocrystallisation molecular features of the solid and the solvent. The formation

can also take place under favourable conditions. According to of new interactions between the molecules of the solid and Accepted previous observations 62,63 , such “favourable conditions” those of the solvent can lead either to partial dissolution of include humidity or the presence of solvent vapour. We found crystallites or at least to increased plasticity of the surface. vapour-assisted cocrystallisation to be an easy method for Both situations would favour the formation of new structures. preparation of pure isoniazid–dicarboxylic acid cocrystals. Figure 6 compares the influence of the solvent on the cocrystal However, the time required for cocrystallisation depended on formation process in solvent-assisted cocrystallisation the length of the dicarboxylic acid. To assess the effect of experiments. From the data presented in Figure 6, it is obvious solvent choice and acid molecule length, we have performed that isoniazid–malonic acid cocrystal and isoniazid–glutaric cocrystallisation experiments in the presence of water, acid cocrystal in the presence of water vapour forms ethanol, acetonitrile, ethyl acetate, chloroform and toluene as considerably faster than other cocrystals. This observation can representatives of polar protic, polar aprotic and non-polar be explained by the exceptionally good water solubility of solvents. The availability of solubility data for pure dicarboxylic malonic and glutaric acids, compared with other dicarboxylic acids in these solvents was also considered as the solubility of acids, implying that good compound solubility can compounds can influence the formation of cocrystals. considerably facilitate vapochemical cocrystallisation. The physical mixtures of isoniazid and dicarboxylic acids were However, vapour-assisted cocrystallisation is possible even if stored in the presence of solvent vapour for 4 h, then the solubility of both compounds is poor. For example,

subjected to PXRD analysis (Figures S24-S31 in ESI). Following isoniazid–dicarboxylic acid cocrystal formation took place in CrystEngComm analysis, samples were exposed to solvent vapour for another the presence of chloroform and toluene vapour, regardless of 20 h and analysed using PXRD (Figures S24-S31 in ESI). Results the low solubility of the compounds in these solvents. of the quantitative analysis of these samples are shown in Generally, the cocrystallisation rate in the presence of organic Figure 5 and in Figure 6. Rates of vapour-assisted solvent vapour follows the order ethanol > ethyl acetate ≈ cocrystallisation reactions show similar trends to those of acetonitrile > chloroform > toluene for all dicarboxylic acids. To mechanochemical cocrystallisation, with the formation of odd- relate the rate of cocrystal formation to properties of these chain acid cocrystals being faster than that of even-chain acid solvents, we compared a set of solvent molecular cocrystals. The reason for the alternation between reaction descriptors 65,66 , such as polarity, dipole moment, dielectric rates is likely related to aforementioned stability differences constant, H-bonding parameters etc. (Table S1 in ESI). This

This journal is © The Royal Society of Chemistry 20xx J. Name ., 2013, 00 , 1-3 | 5

Please do not adjust margins PleaseCrystEngComm do not adjust margins Page 6 of 12

ARTICLE Journal Name

comparison showed that the reaction rate rises with increasing the solvent to form H-bonds can greatly facilitate hydrogen acceptor propensity 65–67 implying that the ability of cocrystallisation. Manuscript

Fig. 5 The amount of the cocrystal in physical mixtures of isoniazid and dicarboxylic acid as a function of acid chain length after storage for 4 and 24 h in the presence of water, ethanol, acetonitrile, ethyl acetate, chloroform and toluene vapour. Accepted CrystEngComm

Fig. 6 The amount of the isoniazid cocrystal with malonic, succinic, glutaric, adipic and pimelic acid in their physical mixtures, depending on the solvent choice.

6 | J. Name ., 2012, 00 , 1-3 This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins Page 7 of 12 PleaseCrystEngComm do not adjust margins

Journal Name ARTICLE

The carboxyl groups of acids are good hydrogen donors and molecule of isoniazid and half a molecule of adipic acid in the therefore, interactions with hydrogen acceptors are likely. asymmetric unit. The relevant crystallographic data are However, the carbonyl oxygen can also act as a hydrogen available in Table S2 in ESI. The adipic acid molecule is located acceptor, making interactions with hydrogen donor on the symmetry centre. Each carboxyl group of the acid forms compounds possible. This is exemplified by the isoniazid– two hydrogen bonds with two distinct isoniazid molecules malonic acid cocrystal, which forms spontaneously, even (Figure 7). The hydroxyl oxygen of the carboxyl group acts as a under ambient conditions (50% RH, 22 °C). Malonic acid has a proton donor to form an O−H···N hydrogen bond to pyridine N

dense hydrogen bonding network, and the p Ka values for both of isoniazid. The carbonyl oxygen, however, is a hydrogen carboxylic groups differ more than those of other diacids 68 . acceptor in a N −H···O hydrogen bond with the isoniazid Consequently, good hydrogen bond donors and acceptors hydrazide group. In addition, two adjacent isoniazid molecules easily interact with malonic acid molecules resulting in a strong form a N −H···O hydrogen bond betw een their hydrazide accelerating effect on the reaction. groups.

Summary of cocrystallisation results To summarize all results obtained in this work, isoniazid–

dicarboxylic acid (C 3–C8 and C10 ) cocrystals could be obtained using mechanochemical, thermal and vapochemical methods. Isoniazid–azelaic acid cocrystal, however, could not be obtained by any of these methods. Vapochemical method failed to provide the formation of the isoniazid–sebacic acid

(C 10 ) cocrystal while mechanochemical and thermal methods Manuscript allowed obtaining this cocrystal. An overview of the isoniazid– dicarboxylic acid cocrystallisation results (whether the cocrystal was obtained by the chosen method) is presented in Table 1. The choice of solvent significantly affects the outcome of vapochemical cocrystallisation; therefore, various solvents should be tested in order to identify the most

appropriate for the system of interest. Fig. 7 Hydrogen bonds in the isoniazid–adipic acid cocrystal polymorph II crystal structure.

Accepted Table 1. Interestingly, the characteristic ring synthons 7 (between the isoniazid pyridine ring and carboxyl group) and 10 The outcome of isoniazid and dicarboxylic acid cocrystallisation experiments using (between hydrazide groups of the isoniazid moieties) common mechanochemical, thermal and vapochemical methods (√ - formation of cocrystal was to isoniazid–dicarboxylic acid cocrystals are not present in this observed). crystal structure. Mechanochemical Thermal Vapochemical Acid method method method Malonic √ √ √ Conclusions Succinic √ √ √ Cocrystallisation of isoniazid to dicarboxylic acids (C –C ) was Glutaric √ √ √ 3 10 conducted using three different cocrystallisation methods: Adipic √ √ √ Pimelic √ √ √ mechanochemical, thermochemical and vapochemical. A comparison of the cocrystallisation results revealed that all Suberic √ √ √ Azelaic – – – three cocrystallisation methods can be used to produce Sebacic √ √ – isoniazid–dicarboxylic acid cocrystals. However, the efficiency of vapochemical and mechanochemical methods depends on the length of the dicarboxylic acid. CrystEngComm Crystal structure of isoniazid–adipic acid cocrystal polymorph II Generally, in mechanochemical and vapochemical reactions The crystal structure of isoniazid–adipic acid cocrystal shorter dicarboxylic acids tend to form cocrystals considerably prepared by crystallisation from methanol has been reported faster than longer ones, partially due to the higher mobility of in the literature 41 . Our mechanochemical and thermochemical small molecules. In all cases cocrystallisation is facilitated by experiments returned a new polymorph (II) of this cocrystal. conditions that promote molecular diffusion (humidity and The crystal structure of the isoniazid–adipic acid cocrystal elevated temperature). polymorph II was solved from powder X-ray diffraction data. The rate of cocrystallisation process was found to show the Similarly to the known polymorph I 41 , the polymorph II odd-even alternation effect in mechanochemical and vapochemical reactions. Cocrystallisation occurred more crystallizes in the monoclinic P21/c space group with a

This journal is © The Royal Society of Chemistry 20xx J. Name ., 2013, 00 , 1-3 | 7

Please do not adjust margins PleaseCrystEngComm do not adjust margins Page 8 of 12

ARTICLE Journal Name rapidly to odd-chain than to even-chain dicarboxylic acids as a dicarboxylic acid. The mixture ( ca . 100 mg) was spread in a result of faster decomposition of the less stable odd-chain acid thin layer on the bottom of a glass vial and the vial was then crystal structures. Vapour-assisted cocrystal formation with placed in a desiccator containing the vapour of water, ethanol, dicarboxylic acids was found to rely largely on the hydrogen- acetonitrile, ethyl acetate, chloroform or toluene. After 4 h of donor and hydrogen-acceptor properties of the solvent. Good storage, PXRD patterns were recorded for the samples. Then proton acceptors offered advantage over other solvents. For these samples were replaced in desiccators containing solvent preparation of pure cocrystal it is important to optimize the vapour and kept for another 20 h. After this storage period crystallisation conditions, such as milling time in PXRD patterns were again recorded for the samples. Three mechanochemistry and solvent choice in vapochemistry. Poor replicates were prepared for each sample and used to choice of these parameters can result in failure of the method calculate the average composition. Samples were not ground to provide the desired product. before PXRD measurements, as grinding can facilitate The presence of the odd-even effect implies that the cocrystallisation. The isoniazid–malonic acid cocrystal was properties of compounds can significantly influence the rate of found to form spontaneously, even under laboratory cocrystal formation. These results show that vapochemical conditions (50% RH, 22 °C) and the isoniazid–malonic acid method is promising for cocrystallisation of small organic molecules and could be used as an alternative method for samples were covered with a polyethylene film during PXRD cocrystal screening. analysis to prevent the effect of moisture.

PXRD analysis Experimental PXRD analysis was performed using a Bruker AXS D8 Advance powder diffractometer (Bruker AXS GmbH, Germany) Materials equipped with a LynxEye position sensitive detector and Cu Kα Isoniazid and dicarboxylic acids were procured from radiation (λ=1.5418 Å), 40 kV, 40 mA. Data were collected at Manuscript commercial suppliers. Prior to cocrystallisation experiments, 5 ambient temperature with a step of 0.02° and scan speed of g of each material were milled in the Retsch MM301 ball mill 0.1 s/step. VT-PXRD experiments to study thermal (10 mL stainless steel jar with one 1 cm stainless steel ball in cocrystallisation of isoniazid and dicarboxylic acids were each) for 10 minutes to reduce the particle size. Milled performed using a Bruker AXS D8 Discover powder samples were kept under ambient conditions for one week diffractometer (Bruker AXS GmbH, Germany) equipped with an and then sieved to obtain the 75 to 150 µm fraction. MRI temperature chamber. Copper Kα radiation (λ=1.5418 Å) was used in the experiments. Diffraction patterns were Cocrystallisation methods recorded with a 0.02° step size and a scan speed of 0.2 s per Thermal cocrystallisation. Thermal cocrystallisation was step in the 2θ range of 3 to 35°. Accepted performed using differential thermal analysis (DTA). Quantitative analysis using Rietveld method Stoichiometric amounts of isoniazid and dicarboxylic acid were 69,70 weighted with a precision of 0.1 mg and gently blended Quantitative Rietveld analyses of the X-ray diffraction data 71 together in an agate mortar. This mixture (6–8 mg) was then were performed using the Bruker Topas 4.2 software with transferred to an aluminium pan. DTA was performed using the fundamental parameters (FP) approach. The crystal 72 the Seiko Exstar6000 TG/DTA6300 (Seiko Instruments Inc., structures of isoniazid (CSD refcode INICAC01), malonic acid Japan). Physical mixtures were heated in open aluminium pans (CSD refcode MALNAC), succinic acid (CSD refcode SUCACB02), at a rate of 5 °C min –1 in a nitrogen flow. glutaric acid (CSD refcode GLURAC), adipic acid (CSD refcode ADIPAC), pimelic acid (CSD refcode PIMELA03), suberic acid Mechanochemical cocrystallisation. Samples for (CSD refcode SUBRAC01), sebacic acid (CSD refcode SEBAAC), mechanochemical cocrystallisation were prepared by isoniazid–malonic acid cocrystal (FADGEY), isoniazid–succinic weighting stoichiometric amounts of isoniazid and dicarboxylic acid cocrystal (FADGIC), isoniazid–glutaric acid cocrystal acid so that the final mass of the sample was 0.3774 grams. (FADGOI), isoniazid–adipic acid cocrystal (FADGUO), isoniazid– Compounds were gently blended together in an agate mortar pimelic acid cocrystal (FADHAV), isoniazid–sebacic acid and placed in 5 mL milling jars containing two 8 mm and three cocrystal (SETROA) were obtained from the Cambridge 6 mm stainless steel balls in each jar (two different milling ball Structural Database (CSD) 73 and used in the calculations. The CrystEngComm sizes were found to offer better data reproducibility). Samples crystal structure of the isoniazid–suberic acid cocrystal was were ground under ambient conditions (45%–55% relative determined by us from single crystal X-ray diffraction data 42 . humidity (RH) and 20–22 °C) for 5 and 15 minutes at a The cell parameters of each structure before quantitative frequency of 20 Hz. Each milling experiment was performed in analysis were refined to PXRD data recorded on the D8 3 replicates. The milling product was analysed by PXRD Advance diffractometer to compensate for the temperature immediately after the experiment. difference during structure determination and PXRD Vapour-assisted cocrystallisation. Physical mixtures for vapour- experiments. The background of the powder patterns was nd assisted cocrystallisation experiments were prepared by gently described by a 2 order Chebyschev polynomial. The blending together stoichiometric amounts of isoniazid and unidentified phases in samples containing malonic and glutaric

Please do not adjust margins Page 9 of 12 PleaseCrystEngComm do not adjust margins

Journal Name ARTICLE

acid were included in the refinement as peak phases. The results obtained were corrected for absorption and sample displacement. The preferred orientation was included in the refinement.

Structure Determination from Powder X-ray Diffraction Data For structure determination PXRD patterns were recorded on a Bruker D8 Discover diffractometer using copper radiation (Cu Kα λ = 1.54180 Å) in transmission mode and a LynxEye (1D) detector. The tube was employed with voltage and current settings of 40 kV and 40 mA. The sample was loaded into a special glass Nr. 10 capillary (0.5 mm diameter). A capillary spinner (60 rpm) was used to minimize instrumental and

sample packing aberrations. Upper knife edge was used to reduce the background produced by air scattering and lower Fig. 8 Final Rietveld fit for isoniazid–adipic acid cocrystal form II: red crosses – knife edge was used to block the primary beam. The incident measured data points; blue line – calculated profile; black line – difference curve; green beam path of the diffractometer was equipped with a Göbel tick marks – calculated peak positions; violet line – background. mirror, Soller slits, and a 0.6 mm divergence slit, while the diffracted beam path was equipped only with Soller slits. The Geometry optimization. Geometry of the isoniazid–adipic acid diffraction patterns were recorded in a 2θ range of 4.5 to 70° cocrystal form II crystal structure was optimized using the DFT at a 0.01° step size with a scan speed of 36 s per step. (Density-Functional Theory) PWscf (Plane-Wave Self- 78 Indexing, space group determination, structure solution and Consistent Field) package within Quantum ESPRESSO .

79 Manuscript Rietveld refinement were conducted using EXPO2014 74 . The Calculations were performed with the PBE exchange unit cell dimensions was determined by the N-TREOR0975 correlation functional, and ultra-soft pseudopotentials with a indexing procedure with a set of 20–25 reflections found in wave function cut-off of 44 Ry and a secondary cut-off of 176 4.5°–40° 2θ range. Space group determination was carried out Ry. Pseudopotentials C.pbe-rrkjus.UPF, F.pbe-n van.UPF, using a statistical assessment of systematic absences, and Z′ N.pbe-rrkjus.UPF, O.pbe-rrkjus.UPF and H.pbe-rrkjus.UPF were was determined based on density considerations. The space acquired from the Quantum ESPRESSO pseudopotential data 80 group was assigned as P2 /c with Z′=1. The cell and diffraction base . Pseudopotentials, energy and force thresholds for 1 81 pattern profile parameters were refined according to the structural relaxation were used as described elsewhere . A k- LeBail algorithm 76 . The background was modelled by a 20 th - point grid of 2 x 2 x 2 was used.

order polynomial function of the Chebyschev type; peak Accepted profiles were described by the Pearson VII function. The initial geometry of molecules was taken from the crystal Notes and references structure of isoniazid–aipic acid cocrystal polymorph I 41 . The 1 J. W. Steed, Trends Pharmacol. Sci. , 2013, 34 , 185–93. simulating annealing technique with dynamical occupancy 2 T. Friščić and W. Jones, J. Pharm. Pharmacol. , 2010, 62 , correction was used to constantly adjust the conformation, 1547–59. position, and orientation of the trial model in the unit cell in 3 N. Qiao, M. Li, W. Schlindwein, N. Malek, A. Davies and G. Trappitt, Int. J. Pharm. , 2011, 419 , 1–11. order to maximize the agreement between calculated and 4 D. Braga, F. Grepioni, G. I. Lampronti, L. Maini and A. measured diffraction data. The Rietveld refinement was Turrina, Cryst. Growth Des. , 2011, 11 , 5621–5627. carried out using soft constraints on bond distances and 5 N. Schultheiss and A. Newman, Cryst. Growth Des. , 2009, 9, angles; In the Rietveld refinement, profile and cell parameters, 2950 – 2967. isotropic thermal vibration, and preferred orientation 6 S. Vijayalakshmi and S. Kalyanaraman, Spectrochim. Acta. parameters 77 were optimized to get an optimum crystal A. Mol. Biomol. Spectrosc. , 2014, 120 , 14–8. structure. The planar pyridine ring was treated as a rigid body, 7 E. D. D’Silva, G. K. Podagatlapalli, S. V. Rao, D. N. Rao and S. and soft constraints on bond distances and angles were used. M. Dharmaprakash, Cryst. Growth Des. , 2011, 11 , 5362– 5369. The U iso values of the H atoms were constrained to be 1.2U iso 8 H. T. Black and D. F. Perepichka, Angew. Chemie - Int. Ed. ,

of those of the non-H atoms. The final Rietveld refinement CrystEngComm showed a good agreement between the observed and 2014, 53 , 2138–2142. 9 A. G. Zalte and R. B. Saudagar, Int. J. Sci. Prog. Res. , 2015, calculated profiles (Figure 8). 12 , 32–35. 10 E. Gagniere, D. Mangin, F. Puel, J. P. Valour, J. P. Klein and O. Monnier, J. Cryst. Growth , 2011, 316 , 118–125. 11 K. Fucke, S. A. Myz, T. P. Shakhtshneider, E. V. Boldyreva and U. J. Griesser, New J. Chem. , 2012, 36 , 1969–1977. 12 T. Leyssens, N. Tumanova, K. Robeyns, N. Candoni and S. Veesler, CrystEngComm , 2014, 16 , 9603–9611. 13 E. Lu, N. Rodríguez-Hornedo and R. Suryanarayanan,

Please do not adjust margins PleaseCrystEngComm do not adjust margins Page 10 of 12

ARTICLE Journal Name

CrystEngComm , 2008, 10 , 665–668. 40 M. Bilde, K. Barsanti, M. Booth, C. D. Cappa, N. M. 14 A. N. Manin, A. P. Voronin, K. V. Drozd, N. G. Manin, A. Donahue, E. U. Emanuelsson, G. McFiggans, U. K. Krieger, Bauer-brandl and G. L. Perlovich, Eur. J. Pharm. Sci. , 2014, C. Marcolli, D. Topping, P. Ziemann, M. Barley, S. Clegg, B. 65 , 56–64. Dennis-Smither, M. Hallquist, Å. M. Hallquist, A. Khlystov, 15 T. Friščić, S. L. Childs, S. A. A. Rizvi and W. Jones, M. Kulmala, D. Mogensen, C. J. Percival, F. Pope, J. P. Reid, CrystEngComm , 2009, 11 , 388–403. M. a. V. Ribeiro da Silva, T. Rosenoern, K. Salo, V. P. 16 S. L. James, C. J. Adams, C. Bolm, D. Braga, P. Collier, T. Soonsin, T. Yli-Juuti, N. L. Prisle, J. Pagels, J. Rarey, A. a. Friščić, F. Grepioni, K. D. M. Harris, G. Hyett, W. Jones, A. Zardini and I. Riipinen, Chem. Rev. , 2015, 115 , 4115–4156. Krebs, J. Mack, L. Maini, A. G. Orpen, I. P. Parkin, W. C. 41 A. Lemmerer, J. Bernstein and V. Kahlenberg, Shearouse, J. W. Steed and D. C. Waddell, Chem. Soc. Rev. , CrystEngComm , 2010, 12 , 2856–2864. 2012, 41 , 413–447. 42 I. Sarcevica, L. Orola, M. V Veidis, A. Podjava and S. 17 T. Friščić, J. Mater. Chem. , 2010, 20 , 7599–7605. Belyakov, Cryst. Growth Des. , 2013, 13 , 1082–1090. 18 N. Shan, F. Toda and W. Jones, Chem. Commun. , 2002, 43 R. Kaur, S. S. R. R. Perumal, A. J. Bhattacharyya, S. 2372–2373. Yashonath and T. N. Guru Row, Cryst. Growth Des. , 2014, 19 S. Karki, T. Friščić and W. Jones, CrystEngComm , 2009, 11 , 14 , 423–426. 470–481. 44 T. J. Chiya and A. Lemmerer, CrystEngComm , 2012, 14 , 20 A. V. Trask, W. D. S. Motherwell and W. Jones, Chem. 5124–5127. Commun. , 2004, 890–891. 45 S. Aitipamula, A. B. H. Wong, P. S. Chow and R. B. H. Tan, 21 A. V Trask, N. Shan, W. D. S. Motherwell, W. Jones, S. Feng, CrystEngComm , 2013, 15 , 5877–5887. R. B. H. Tan and K. J. Carpenter, Chem. Commun. , 2005, 46 S. M. A. Mashhadi, U. Yunus, M. H. Bhatti and M. N. Tahir, 880–882. J. Mol. Struct. , 2014, 1076 , 446–452. 22 H. Yamashita, Y. Hirakura, M. Yuda, T. Teramura and K. 47 N. Ravikumar, G. Gaddamanugu and K. Anand Solomon, J. Terada, Pharm. Res. , 2013, 1, 70–80. Mol. Struct. , 2013, 1033 , 273–279. 23 C. Maheshwari, A. Jayasankar, N. A. Khan, G. E. Amidon 48 I. Sarcevica, L. Orola, M. V Veidis and S. Belyakov, Acta and N. Rodríguez-Hornedo, CrystEngComm , 2009, 11 , 493– Crystallogr. Sect. C, Struct. Chem. , 2014, 70 , 392–395. Manuscript 500. 49 B. Swapna, D. Maddileti and A. Nangia, Cryst. Growth Des. , 24 K. K. Arora, N. G. Tayade and R. Suryanarayanan, Mol. 2014, 14 , 5991–6005. Pharm. , 2011, 8, 982–989. 50 A. Lemmerer, J. Bernstein and V. Kahlenberg, 25 A. Jayasankar, D. J. Good and N. Rodríguez-Hornedo, Mol. CrystEngComm , 2011, 13 , 5692–5708. Pharm. , 2007, 4, 360–372. 51 M. Sevukarajan, R. Sodanapalli, B. Thanuja and R. Nair, J. 26 D. Braga, S. L. Giaffreda, F. Grepioni, M. R. Chierotti, R. Biomed. Sci. , 2011, 3, 397–402. Gobetto, G. Palladino and M. Polito, CrystEngComm , 2007, 52 P. Grobely, A. Mukherjee and G. R. Desiraju, 9, 879–881. CrystEngComm , 2011, 13 , 4358–4364. 27 D. Braga, G. Palladino, M. Polito, K. Rubini, F. Grepioni, M. 53 A. Lemmerer, CrystEngComm , 2012, 14 , 2465–2478.

R. Chierotti and R. Gobetto, Chem. - A Eur. J. , 2008, 14 , 54 Z.-J. Jiang, L. Lu-De, X.-D. Wu and X.-J. Yang, Chinese J. Accepted 10149–10159. Inorg. Chem. , 2009, 25 , 746–750. 28 A. Jayasankar, A. Somwangthanaroj, Z. J. Shao and N. 55 M. C. R. Freitas, J. M. S. António, R. L. Ziolli, M. I. Yoshida, Rodríguez-Hornedo, Pharm. Res. , 2006, 23 , 2381–2392. N. A. Rey and R. Diniz, Polyhedron , 2011, 30 , 1922–1926. 29 R. Bobrovs, L. Seton and N. Dempster, CrystEngComm , 56 J. C. Hanson, N. Camerman and A. Camerman, J. Med. 2015, 17 , 5237–5251. Chem. , 1981, 24 , 1369–71. 30 L. J. Thompson, R. S. Voguri, L. Male and M. Tremayne, 57 G. G. Tsintsadze, G.V. Dzhavakhishvili, Z.O. Aleksandrov, Y. CrystEngComm , 2011, 13 , 4188–4195. T. Struchkov and A. P. Narimanidze, Russ. J. Coord. Chem. , 31 C. B. Aakeröy, S. Forbes and J. Desper, J. Am. Chem. Soc. , 1980, 6, 785. 2009, 131 , 17048–17049. 58 L. B. Zinner, D. E. Crotty, T. J. Anderson and M. D. Glick, 32 D. Braga, E. Dichiarante, G. Palladino, F. Grepioni, M. R. Inorg. Chem. , 1979, 18 , 2045–2048. Chierotti, R. Gobetto and L. Pellegrino, CrystEngComm , 59 L. BenHamada and A. Jouini, Mater. Res. Bull. , 2006, 41 , 2010, 12 , 3534 – 3536. 1917–1924. 33 P. Vishweshwar, A. Nangia and V. M. Lynch, Cryst. Growth 60 V. O. Gel’mbol’dt, V. N. Davydov, L. V. Koroeva and E. V. Des. , 2003, 3, 783 – 790. Ganin, Russ. J. Inorg. Chem. , 2002, 47 , 987. 34 V. R. Thalladi, M. Nüsse and R. Boese, J. Am. Chem. Soc. , 61 V. S. Mandala, S. J. Loewus and M. A. Mehta, J. Phys. Chem. 2000, 122 , 9227–9236. Lett. , 2014, 5, 3340–3344. 35 H. Zhang, Q. Yin, Z. Liu, J. Gong, Y. Bao, M. Zhang, H. Hao, 62 T. Ervasti, J. Aaltonen and J. Ketolainen, Int. J. Pharm. , B. Hou and C. Xie, J. Chem. Thermodyn. , 2014, 77 , 91–97. 2015, 486 , 121–130. CrystEngComm 36 H. Zhang, C. Xie, Z. Liu, J. Gong, Y. Bao, M. Zhang, H. Hao, B. 63 I. Sarceviča, L. Orola, K. P. Nartowski, Y. Z. Khimyak, A. N. Hou and Q. Yin, Ind. Eng. Chem. Res. , 2013, 52 , 18458– Round and L. Fábián, Mol. Pharm. , 2015, 12 , 2981–2992. 18465. 64 A. Newman and G. Zografi, J. Pharm. Sci. , 2014, 103 , 1–10. 37 M. Bilde, B. Svenningsson, J. Mønster and T. Rosenørn, 65 C.-H. Gu, H. Li, R. B. Gandhi and K. Raghavan, Int. J. Pharm. , Environ. Sci. Technol. , 2003, 37 , 1371–1378. 2004, 283 , 117–25. 38 E. A. Bruns, J. Greaves and B. J. Finlayson-Pitts, J. Phys. 66 D. Xu and N. Redman-Furey, Int. J. Pharm. , 2007, 339 , 175– Chem. A , 2012, 116 , 5900–5909. 188. 39 C. D. Cappa, E. R. Lovejoy and A. R. Ravishankara, J. Phys. 67 C. A. Hunter, Angew. Chemie - Int. Ed. , 2004, 43 , 5310– Chem. A , 2007, 111 , 3099–3109. 5324.

Please do not adjust margins Page 11 of 12 PleaseCrystEngComm do not adjust margins

Journal Name ARTICLE

68 C. Tong, M. Blanco, W. A. Goddard and J. H. Seinfeld, Environ. Sci. Technol. , 2004, 38 , 3941–3949. 69 H. M. Rietveld, Aust. J. Phys. , 1988, 41 , 113–116. 70 D. L. Bish and S. A. Howard, J. Appl. Crystallogr. , 1988, 21 , 86–91. 71 R. W. Cheary and A. Coelho, J. Appl. Crystallogr. , 1992, 25 , 109–121. 72 T. N. Bhat, T. P. Singh and M. Vijayan, Acta Crystallogr. Sect. B Struct. Sci. , 1974, 30 , 2921–2922. 73 F. H. Allen, Acta Crystallogr. Sect. B Struct. Sci. , 2002, 58 , 380–388. 74 A. Altomare, C. Cuocci, C. Giacovazzo, A. Moliterni, R. Rizzi, N. Corriero and A. Falcicchio, J. Appl. Crystallogr. , 2013, 46 , 1231–1235. 75 A. Altomare, G. Campi, C. Cuocci, L. Eriksson, C. Giacovazzo, A. Moliterni, R. Rizzi and P.-E. Werner, J. Appl. Crystallogr. , 2009, 42 , 768–775. 76 A. Le Bail, H. Duroy and J. L. Fourquet, Mater. Res. Bull. , 1988, 23 , 447–452. 77 W. A. Dollase, J. Appl. Crystallogr. , 1986, 19 , 267–272. 78 P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, G. L. Chiarotti, M. Cococcioni, I. Dabo, A. Dal Corso, S. de Gironcoli, S. Fabris, G. Fratesi, R. Gebauer, U. Gerstmann, C. Gougoussis, A. Kokalj, M. Lazzeri, L. Martin-Samos, N. Marzari, F. Mauri, R. Manuscript Mazzarello, S. Paolini, A. Pasquarello, L. Paulatto, C. Sbraccia, S. Scandolo, G. Sclauzero, A. P. Seitsonen, A. Smogunov, P. Umari and R. M. Wentzcovitch, J. Phys. Condens. Matter , 2009, 21 , 395502. 79 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett. , 1996, 77 , 3865–3868. 80 http://www.quantum-espresso.org/pseudopotentials . 81 A. M. Lund, A. M. Orendt, G. I. Pagola, M. B. Ferraro and J. C. Facelli, Cryst. Growth Des. , 2013, 13 , 2181–2189.

Accepted CrystEngComm

Please do not adjust margins CrystEngComm Page 12 of 12

For Table of Contents only:

Systematic study on mechanochemical, thermal and vapochemical cocrystallisation demonstrates the effect of compound properties on the outcome of the reaction.

Manuscript Accepted CrystEngComm